Hi Sean,

On 11:35 Thu 29 Jan     , Sean Hefty wrote:
> 
> Stan came across an issue while porting saquery to Windows resulting from
> duplicate definitions for IB_MAD_METHOD_*.  These are defined as enums in 
> mad.h,
> but as #define values in ib_types.h.

Why it should be a problem? ib_types.h is included in saquery.c after
mad.h and as far as I understand "defines" will work instead of "enums".
Opposite inclusion order could cause the problem of course.

> Do you have some preferred way to remove
> the duplicate defines? Can ib_types.h include mad.h and just remove the
> duplicates?

It is not an easy - libibmad and opensm are not dependent packages. Both
have a reach history ( :( ) and external usage. Introducing dependency
here would be noisy thing. And it is not really needed for any other tool
(I think saquery.c is only exception where mad.h and ib_types.h are used
together).

Sasha
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Reply via email to